Richard K. Wilson, director of the Genome Institute at Washington University in St. Louis, calls the move to genomic-based studies a paradigm shift.

The studies are part of the Cancer Genome Atlas project, an initiative funded by the National Institutes of Health to map genetic information from 10,000 tumors covering some two dozen cancers. The project has already yielded findings in other cancers, including breast, lung and colon that are helping to fundamentally alter medicine's views about the western world's leading killer.

"Cancer is a disease of the genome," said Brad Ozenberger, program director for the Atlas studies at the National Human Genome Research Institute, which along with the National Cancer Institute is spearheading the project. With recent advances in sequencing technology, scientists "are able to really dive in deep and find out what is causing the disease."

Cancer has long been identified and treated based on the organ where it arises—breast, prostate, skin, for instance. But emerging genomic information shows that "a lot of these different cancers share a similar genetic identity," said Richard K. Wilson, director of the Genome Institute at Washington University in St. Louis. "We started…thinking that we should treat cancer based not on what tissue it comes from but on what genes are mutated."

That insight is prompting drug makers to develop drugs that target specific mutations in patients' tumors. A new term is emerging to describe this approach: precision medicine.

Such an approach may ultimately require multiple agents to attack all genetic drivers of a tumor, which can manage to eventually find ways around a single effective drug. And such therapies are proving to be expensive. But cancer researchers are convinced that genomic-based strategies offer a chance to make major progress against the disease.

Dr. Wilson called it a paradigm shift: "The more cancer genomes we look at the more that is reinforced."

Some 50,000 new cases of endometrial cancer are diagnosed in the U.S. each year, with about 8,000 deaths. Current methods have identified two tumor types. Type I tumors are typically treated with radiation after the cancer is surgically removed, and the prognosis for recovery is good. Type II tumors are considered more aggressive, requiring surgery plus chemotherapy, and are associated with poorer outcomes.

But a genomic analysis of 373 endometrial tumors by Atlas researchers identified four subtypes instead of two, each possibly requiring different treatments. In addition, the scientists found that about 25% of tumors that pathologists classified as high-grade type I had molecular traits similar to type II, suggesting the need for more aggressive treatment.

The findings need to be validated in clinical trials, said Douglas A. Levine, a gynecologic oncologist who leads the Gynecology Research Laboratory at Memorial Sloan-Kettering Cancer Center. But "we're now going to put more patients into the group that should get more aggressive treatment after surgery," he said. The endometrial cancer findings were published by the journal Nature.

The study also found that a trait involving a gene called PI3K was more common in endometrial tumors than in any other tumor type studied so far by the Atlas project. Several drug companies have drugs targeting PI3K in development, raising prospects that drug trials could soon be launched testing such therapies for endometrial cancer.

About 15,000 new cases of acute myeloid leukemia, or AML, are diagnosed each year and most such patients end up dying of the disease. It is a cousin to the less aggressive chronic myeloid leukemia that helped usher in the era of targeted therapy when the genetic mutation driving the disease responded dramatically to the Novartis AG NOVN.VX -0.07% drug Gleevec approved in 2001.

In AML, "we don't see a simple smoking gun like you see in CML," said Washington University's Dr. Wilson. At least two of the 200 AML patients whose tumors were studied in the Atlas project had some 260 different mutations associated with their cancer, he said. "That's pretty amazing. It's a much more complicated genetic picture than we might have expected."

Still, he said the genomic analysis points to mutations that correlate with aggressive forms of AML. "If you know that up front, you can be more aggressive in treating a patient," Dr. Wilson said. It also might call for considering a bone-marrow transplant much earlier, providing time "to identify the perfect donor."

AML often strikes the elderly. A genetic profile indicating a less-aggressive form of the disease might enable doctors to spare frail patients "heavy duty chemo that is almost worse than the disease," he said.

Findings from the AML study were published by the New England Journal of Medicine.